yet none was as slight as
those of Mr. von Osten.[I] I further noticed that Count Matuschka and
Mr. Schillings often showed a tendency to accompany every tap of the
horse with a slight nod of the head, the last being accompanied by a
more pronounced nod and then followed by the upward jerk of the head, in
other words, they beat time with the horse. In the case of the last
three mentioned, for whom the horse responded far less effectively than
for Mr. von Osten or Mr. Schillings, belated or precipitate jerks would
frequently occur. This was found to be true in the case of all other
persons who had failed to elicit adequate responses from the horse.
Often, in both cases, a complete absence of any kind of minimal movement
had been noted. The accuracy of these observations in the case of Mr.
von Osten is attested by Mr. Stumpf and Mr. von Hornbostel, and by these
same gentlemen and Prof. F. Schumann in the case of Mr. Schillings and
myself. They also found these movements to be most minute in the case of
Mr. von Osten. In my case also they pronounced them "minimal, and often
quite imperceptible". All other persons who have seen me work with the
horse, but who were not familiar with the nature of these movements,
never perceived them, no matter how closely they observed me.

[Footnote I: During the tests Mr. von Osten nearly always wore a
slouch hat with a wide rim. The rim, of course, always moved with
the head, and made the movements appear on a larger scale, (in the
ratio of about 3:2, as I was able to ascertain later by graphic
methods). But observation was successful, even at a distance of a
meter and a half, when he worked with head uncovered. And even if
head and forehead were covered entirely, it was still possible to
note the movements by watching the eye-brows. When Mr. Schillings
and the rest of us worked with the horse, we either went bare-headed
or wore only a very small cap.]

Since the doubt was expressed that these movements did not precede but
followed closely upon the back-step of the horse (i. e., that an error
with regard to the time-element was involved), it became important that
time measurements be taken. This was done in the following manner: The
questioner asked the horse to tap numbers from 5 to 20, seldom higher.
He purposely refrained from pronouncing the number, but recorded it
after each test had been completed. This was a matter of indifference to
the ho